Rights-infringing copies of “NEINhorn”: Carlsen sues OpenAI | heise online

heise+ entdecken

SuchenAbo

Suchen

Alle Magazine im Browser lesen<br>Newsletter<br>heise-Bot<br>Push-Nachrichten

${lead}

${lead}

${content}

${content}

${content}

${content}

Advertisement

Advertisement

Carlsen Verlag has filed a lawsuit against OpenAI because the AI chatbot ChatGPT generates stories of the copyrighted “NEINhorn” upon request and was apparently trained “in an unlawful manner” with the children’s books. The Hamburg-based publisher made this public on Thursday and criticized that the illustrations generated by ChatGPT could hardly be distinguished from the original by author Marc-Uwe Kling and illustrator Astrid Henn. The AI chatbot even makes “proposals on its initiative for further infringing texts and illustrations.” These also include characters and settings not mentioned. ChatGPT even offers “complete print templates for illustrated stories, including cover design, imprint, a fake ISBN, and publisher’s logo”.

Continue after ad

“Steal and get away with it?”

Given the “extremely detailed copies”, it must be assumed that the original works from the children’s book series “were used unlawfully for training OpenAI’s language and image models and are now contained as memorization in these models, so that they can be accessed at any time via ChatGPT,” writes Carlsen. OpenAI is being asked to cease the unlicensed use of protected works, disclose any irregularities, and take full responsibility for illegal actions, explains Christian Schumacher-Gebler, CEO of the German Bonnier Group, to which Carlsen belongs. In particular, any damage incurred must be compensated. The lawsuit was filed at the Munich I Regional Court, which is not unfamiliar to OpenAI.

“Anyone who illegally downloads even a single film, book, or audiobook faces a severe penalty. But OpenAI simply steals all the art and culture of the whole world and thinks it can get away with it?” asks Marc-Uwe Kling in the statement. His little unicorn would certainly have a clear answer to that. Astrid Henn points out that her illustrations are often the result of months of work. If they are used without permission for training AI, “with the goal that the AI makes my work, my property, usable for everyone,” then that is theft. For her, that is no different from stolen company secrets being sold on the darknet.

Videos by heise

mehr Videos

c't 3003

heise & ct

Peertube

The NEINhorn, only seven years old, is already a children’s book classic. The book of the same name humorously tells how important it is to say no. There are now two more parts featuring other characters known for certain words, including the KönigsDOCHter, the NAhUND, and the WASbär. In addition, the drawings have a high recognition value, which is said to be particularly evident in the AI-generated imitations. For Carlsen, the debate about AI and children’s books is not new; a year ago, the publisher clarified that it would take legal action against “inhuman, racist, violence-glorifying, and pornographic uses of the Conni character” in AI memes that were popular at the time.

(mho)

Don't miss any news – follow us on<br>Facebook,<br>LinkedIn or<br>Mastodon.

This article was originally published in

German.

It was translated with technical assistance and editorially reviewed before publication.

Dieser Link ist leider nicht mehr gültig.

Links zu verschenkten Artikeln werden ungültig,<br>wenn diese älter als 7 Tage sind oder zu oft aufgerufen wurden.

Sie benötigen ein heise+ Paket, um diesen Artikel zu lesen. Jetzt eine Woche unverbindlich testen – ohne Verpflichtung!

Wochenpass bestellen

Sie haben heise+ bereits abonniert?

Hier anmelden.

Oder benötigen Sie

mehr Informationen zum heise+ Abo

Anzeige

Advertisement

In fünf Monaten zum kompletten Systemwechsel

Dell XPS: Zwei Jahrzehnte Leistung und Design

XPS: Die Geschichte hinter den drei Buchstaben

Infrastruktur in der KI-Ära: Netz rückt in die erste Reihe

Mit UEM zu mehr IT-Effizienz

Themenspecial: Digitale Souveränität

Warum Unternehmen Instagram nutzen sollten

Sensible Daten schützen

Cyber-Hygiene statt Cyber-Panik

BSI-Vorgaben für Cloud-Souveränität

Shortlink:

https://heise.de/-11420677

Advertisement

Advertisement

Newsletter

heise-Bot

heise-Bot

Push Nachrichten

Push

Push-Nachrichten

kopieren